Image Alignment in ElectroSpotmatic

What It Does

Image Alignment automatically compensates for camera movement between frames when capturing computational photography sequences. Even with a tripod, slight camera shake or movement can cause misalignment. This feature ensures all images in a sequence are perfectly registered before they're combined.

Image alignment is automatically performed as the first step in focus stacking, motion removal, HDR merging, long exposure, and astrophotography workflows.

You can download test images for use with SpotmaticMagic from: Google Drive 

When Alignment Is Used

Image alignment is automatically applied in:

• Focus Stacking: Compensates for "focus breathing" (magnification changes as focus shifts)
• Motion Removal: Ensures frames are registered so moving objects can be accurately detected
• HDR Merging: Aligns exposure-bracketed images to prevent ghosting artifacts
• Long Exposure: Aligns frames before averaging to prevent blur from camera movement
• Astrophotography Stacking: Aligns star fields across multiple frames

How to Configure Alignment Settings

Configure in iOS Settings → ElectroSpotmatic → Image Alignment Algorithm:

• Vision Framework (Recommended): Fast, reliable, memory-efficient
• IC-LM: High-precision affine alignment for complex transformations
• Metal ECC: GPU-accelerated (experimental)

Alignment Methods Explained

Vision Framework (Default)

Best for: Most photography workflows

Advantages: Fast processing, memory efficient, robust to intensity changes, reliable results

Limitations: Only handles translation (no rotation, scale, or shear)

IC-LM (Inverse Compositional Levenberg-Marquardt)

Best for: Macro photography, telephoto focus stacking, scenes with rotation/shear

Advantages: Handles complex transformations (6-DOF affine), sub-pixel precision, excellent for focus breathing compensation

Limitations: Slower processing (1.5-2× slower), higher memory usage

Metal ECC (Enhanced Correlation Coefficient)

Best for: Experimental use, testing

Limitations: High memory usage, may fail to converge, slower than Vision Framework

Best Practices

• Use a tripod for better results and to preserve more image area
• Keep camera position consistent between frames
• Ensure consistent lighting across frames
• Scenes with good detail and texture align better
• Leave on Vision Framework for most cases, switch to IC-LM only for macro/telephoto work

Troubleshooting

Alignment takes a long time: IC-LM is slower - switch to Vision Framework unless you specifically need IC-LM

Aligned images are cropped: This is expected - images are cropped to the common intersection after alignment

Poor alignment results: Ensure scene has sufficient texture, check that images are from the same scene, try switching to IC-LM

Alignment fails: Very featureless scenes may not align well, ensure there's at least 30-40% overlap between frames